You may also find (my) [ANN] FlexiJoins.jl: fresh take on joining datasets package useful.
- Lots of join conditions available, including interval overlap
- Supports a wide range of collections/table types, even
DataFrames
- Uniform interface and composability, as in
by_key(:name) & by_pred(:value, ∈, :valrange)
- Lots of other neat features – see the linked thread and the examples notebook.
- Optimized, as in not looping over
N*M
pairs. Still, very specialized algorithms can be somewhat faster.